Will next 3dmark's game test 4 require sm 3.0?

[Nick[FM]]

come on guys! Boozembly is still 2 months away.

(I'll leave to worm to explain what is a Boozembly. I have a such bad alcohol tolerance (and it's even known in pretty many places, thanks to few beers in emptines and a waaaaayyy too long ICQ list ) that I might get drunk even talking about it...)

Hehe, "Boozembly" is the annual behind the scene party at the Assembly, where all the old farts from the Amiga, C64 and perhaps nowdays also PC demo-scene gather to .. well, get drunk. Very, very drunk. I must admit that I have only visited the Boozembly once or twice, and been completely sober.

Sorry for completely OT stuff.

 

[Evildeus]

BTW Worm, any idea when (and if ) we will see some next approuved driver from Nv?

 

[Simon F]

I think I'll take the 5'th on this poll.

So when did you become a US citizen?

 

[Hanners]

BTW Worm, any idea when (and if ) we will see some next approuved driver from Nv?

Well, nVidia will have to actually release a driver first...

 

[Evildeus]

BTW Worm, any idea when (and if ) we will see some next approuved driver from Nv?

Well, nVidia will have to actually release a driver first...

Well, i suppose FM have them before

 

[WaltC]

I really wish polls of this sort would recognize the value of adding a "Don't know" or "Maybe" category, as that's how I would have answered it, as I think what will determine the probability of any ps3.0 software support of ps3.0 features beyond ps2.0b+, is whether or not the nV40 implementation of those features is functional and worthwhile (likewise with ATi's ps2.0+ feature support above R3x0's ps2.0.)

Certainly, any sort of "we haven't exposed them in the drivers yet" posture from nVidia is the same as to say they aren't there at all, as their efficacy cannot be demonstrated, and unexposed hardware functionality is equivalent to absent hardware functionality. The assumption behind such questions is that the nV40 implementation of those features is such that software developers would see utility in supporting them, but such is not the case at this time since their application in nV40 has yet to be publicly demonstrated, even by nVidia.

So I would answer "Maybe," as I think it depends on the quality, or lack thereof, of the nV40 ps3.0 feature-set implementation.

 

[Reverend]

"NVIDIA paid extra money to FM for next 3DMark to include SM 3.0 usage altho FM didn't have the original intention"

Now, that can be a likely post by some IHV fans.

IF there is indeed to be some 3.0 in the next 3DM.

 

[digitalwanderer]

NVIDIA paid extra money to FM for next 3DMark to include SM 3.0 usage altho FM didn't have the original intention

Page 1 baby! 8)

 

[Reverend]

Huh?

 

[digitalwanderer]

Huh?

Sorry, "Page 1 baby! 8)", could be roughly translated into, "I'm posting that quote of yours up on me frontpage and am gonna pimp that baby for all it's worth all up and down the intraweb!"

I hope that clarifies things.

 

[MrBond]

I guess it depends. If 3dm is indeed supposed to be a "forward" looking benchmark, then it would be good to know what games are going to be doing in the future. At least from this standpoint, it looks like PS3.0 is going to be nearly skipped over (with the exception of TWIMTBP devs and ones who don't mind spending a little more time coding for it) in favor of whatever the Next big thing after 2.x is

If 3dmark can do multiple code paths on GT4, then I think they should do it. After all, we need a good, satisfying synthetic test to stress out the 6800, and see how good the PS3 vs PS2 implementation is after all (and put those "PS3.0 unplayable on 6800 rumors to rest"). Of course, they probably shouldn't make it a PS3.0 only test, as that would rile the suspicions of ATI fans and most games should be using HLSL to compile for 2.0, 2.x, and 3.0 anyway.

 

[digitalwanderer]

Rev-Oh, and I really meant to add a winky to the end of both of them to let ya know I was joking....but evily spaced it for a few minutes just to keep ya guessing.

 

[KimB]

Well, there are a number of different added features that may or may not get used in games, due to a number of different reasons. If, for example, floating-point texture filtering and blending added extra performance penalties beyond the obvious memory size and bandwidth penalties, it may not be used (which I sure hope not. This would be one of my first tests when I finally get my hands on one, if somebody else doesn't get to it first).

P.S. I can't believe I made that typo. I typed "nads" instead of "hands" ... ugh...

 

[digitalwanderer]

I can't believe I made that typo. I typed "nads" instead of "hands" ... ugh...

Your freudian slip is showing...now you'll never be able to convince me that your PC ain't a phallic symbol for ya.

 

[rainz]

3Dmark05 will be release prolly around october or november 2004 ... And i dont expect a lot from it .. Even the current gaming computers painfully run the Game Nature test in the 2003 .. I dont think we'll see a big improvement in the graphics. Well that's what i think. ( Yeh prolly a difference .. but not as huge as the previous 2001 .. 2001 VS 2003 )

RainZ

 

[KimB]

Your freudian slip is showing...now you'll never be able to convince me that your PC ain't a phallic symbol for ya.

I don't know. It just sounds painful....

 

[pat777]

]

I would be surprised if the next 3D Mark didn't require SM3.0 for Test 4. I expect it would be something like this:
Test 1) SM1.1+
Test 2) SM2.0+
Test 3) SM2.0+
Test 4) SM3.0

Of course it's always possible that Test 4 might need only VS3.0 which could be done in software.

If you spotted our April newsletter, it might give you at least a small indication what comes to Shaders below 2.0.. If you didn't, here's a smallish quote from it:

I don't think this first bit of info will come as a big surprise to most of you, but the Next 3DMark will require a fully DirectX 9 compliant graphics card capable of at least PixelShader 2.0.

3dmark2003 didn't require PS 2.0 for testing, you just don't test game 4. You still can publish 3dmark2003 scores with a DX8 card.

 

[KimB]

I don't think anybody expects the next 3DMark to require PS3 for all tests. The question is whether or not it will be required for one test that contributes to the score.

 

[Mariner]

I don't think anybody expects the next 3DMark to require PS3 for all tests. The question is whether or not it will be required for one test that contributes to the score.

Personally, I hope the equivalent of "Game Test 4" will have some support for SM3.0 - as long as they make sure the test can also be run on SM2.0 cards through a multipass/whatever fallback mode. They also need to make sure that the weighting of the score is fair - the majority of the score needs to be relating to PS2.0 performance, especially when you consider that low-end chips with decent PS2.0 are just beginning to be released. The NV4X budget series aren't expected for months yet are they and I don't doubt that their PS2.0 performance will be pretty good in relation to PS3.0.

 

[Nick[FM]]

]

I would be surprised if the next 3D Mark didn't require SM3.0 for Test 4. I expect it would be something like this:
Test 1) SM1.1+
Test 2) SM2.0+
Test 3) SM2.0+
Test 4) SM3.0

Of course it's always possible that Test 4 might need only VS3.0 which could be done in software.

If you spotted our April newsletter, it might give you at least a small indication what comes to Shaders below 2.0.. If you didn't, here's a smallish quote from it:

I don't think this first bit of info will come as a big surprise to most of you, but the Next 3DMark will require a fully DirectX 9 compliant graphics card capable of at least PixelShader 2.0.

3dmark2003 didn't require PS 2.0 for testing, you just don't test game 4. You still can publish 3dmark2003 scores with a DX8 card.

Oh, I was refering to the Next 3DMark, not 3DMark03. Sorry for any confusion. 3DMark03 can be run even with a DX7-class card (GT1 only though).